In 1939, Rackemann and Greene reported a subgroup of patients with polyarteritis nodosa and concomitant allergic disease. Similar findings were reported in the early 1940s by Harkavy. The histopathology and clinical features associated with this disease entity were first described in 1951 by Churg and Strauss, who reported a form of necrotizing vasculitis in several organs, associated with eosinophilic tissue inflammation and extravascular granulomas, occurring in asthmatics, with associated fever and peripheral hypereosinophilia (Churg–Strauss syndrome).138 This disease entity, now termed eosinophilic granulomatosis with polyangiitis (EGPA), is an uncommon systemic disease.139–141 An overall prevalence of 10.7 to 13 cases per million inhabitants is estimated among the general population,142 and up to 64 cases per million persons per year is estimated among patients with a history of asthma. The mean annual incidence has been estimated at 0.5 to 6.8 per million population across various countries.142
Approximately 10% of all patients with vasculitis prove ultimately to have EGPA (see also Chapter 74). Nevertheless, the precise incidence of EGPA is unknown due to uncertainties regarding diagnosis and variable clinical presentation. The true incidence of EGPA may be higher than is generally recognized, since the syndrome has many clinical, radiographic, and histologic features in common with other vasculitic, eosinophilic, and granulomatous disease states. The diagnosis of EGPA may be missed if not carefully entertained.
EGPA may occur in patients of any age, but it develops most commonly in patients between the ages of 38 and 50.1,143,144 It is rare in persons older than age 65, as well as children and adolescents. However, among affected children, it tends to follow a more aggressive clinical course.145 There is no clear gender predominance.142,144 Among women, disease onset has been reported during pregnancy. The MHC complex DRB4 allele may confer genetic risk for the disease.146,147
EGPA tends to follow a subacute course, with symptoms ranging over months to years. Historically, three distinct clinical phases of the disease have been recognized: the prodromal phase, the eosinophilic phase, and the vasculitic phase.148,149 The prodromal phase has been characterized by “late-onset” (in the second or third decade) allergic rhinitis and atopy in persons often lacking a family history of atopy. Severe allergic rhinitis, sinusitis, and drug sensitivity are usually present for 8 to 10 years, and up to 30 years before EGPA disease recognition. Asthma is a feature of EGPA in all cases. It typically precedes the onset of vasculitis by 3 to 6 years.148 The eosinophilic phase is typified by the development of marked peripheral blood eosinophilia and eosinophilic tissue infiltration, most commonly of the lung, GI tract, and skin. The vasculitic phase is characterized by vasculitis of the small and medium vessels with vascular and extravascular granulomas. The onset of the vasculitic phase is often heralded by development of constitutional symptoms, including fever, malaise, weight loss, arthralgias, myalgias, and increased allergic or asthmatic symptoms. Although the vasculitis tends to occur several years after the onset of allergic manifestations of the disease, in some cases it develops within months of, or concomitant with, the onset of asthma. A short duration between the onset of asthma and vasculitis has been associated with increased severity of vasculitis. During the vasculitic stage, the asthma symptoms may persist and worsen, or they may diminish. When asthma dissipates, it often flares later in the course of illness and may require prolonged steroid treatment. Importantly, not all patients progress sequentially through these phases, and overlap of clinical features between phases is often present. Although EGPA typically affects multiple organ systems, limited forms of disease have also been described.142,150 Manifestations in the lungs, heart, skin, and nervous system are most common.
Lung involvement occurs in nearly all patients with EGPA.142 Most of the respiratory manifestations of EGPA occur in the prodromal and eosinophilic phases of the disease. As noted earlier, all patients have asthma at some point in the illness. Upper airway allergic disease, including sinusitis, rhinitis, and polyposis, is seen in 75% to 85% of patients and may be the presenting symptom.1,142 Unlike granulomatosis with polyangiitis, necrotizing granulomas involving the upper airway are unusual in EGPA. The asthma and upper airway disease usually are long-standing, severe, and often require steroid therapy (systemic or inhaled) to maintain control of symptoms.142 Spirometry may reveal an obstructive ventilatory defect.151 A Loeffler-like syndrome with eosinophilic infiltration of the lung parenchyma is seen in 50% to 70% of patients. These patients may develop dyspnea, cough, and wheezing. Their chest radiographs have transient, migratory, patchy, nonlobar, nonsegmental, often peripheral pulmonary infiltrates, with no regional predilection.152–154 Nodular lesions, reticular opacities, bronchial wall thickening, bronchiectasis, and hilar adenopathy are less common findings. In contrast to granulomatosis with polyangiitis, the allergic granulomas cavitate more rarely. Up to 50% of patients develop unilateral or bilateral pleural effusions,154 which may be associated with pleuritic chest pain. The chest radiograph may occasionally be normal. HRCT scanning has demonstrated bronchial wall thickening, pulmonary artery enlargement (in comparison to the corresponding bronchi), irregular stellate configuration of some vessels, adenopathy, areas of interlobular septal thickening, and scattered patchy parenchymal opacities with ground-glass, nodular, consolidated, or tree-in bud appearance.1 Nodules may be present within areas of ground glass.90,154,155 These findings have been reported to correlate with pathologic findings evident on open lung biopsy such as eosinophilic pneumonia, alveolar hemorrhage, eosinophilic infiltration of the bronchial wall, and septum.153 Further studies are necessary to determine whether high-resolution CT is useful to stage the disease or establish the diagnosis without tissue biopsy.
Cardiac manifestations generally are not evident on initial presentation of EGPA. However, they typically occur during the vasculitic phase of the disease and are a major source of morbidity and the principal cause of death (in up to 50% of cases) from the disorder.142,143,156 Patients may be asymptomatic. Progressive congestive heart failure (CHF) occurs in up to 47% of cases because of endomyocardial infiltration by eosinophils1,157 or ischemic cardiomyopathy resulting from necrotizing vasculitis of the coronary arteries.158 This coronary vasculitis is fatal up to 60% of the time. Acute pericarditis is present in approximately one-third of cases, and cardiac tamponade has been reported. Constrictive pericarditis may develop over time. Cardiac involvement is more common in persons with EGPA who lack serum ANCA (see below).159,160
A wide array of neurologic manifestations may develop in EGPA. Mono- or polyneuropathy (most notably mononeuritis multiplex) is present in 69% to 75% of cases. The common peroneal, ulnar, and internal popliteal nerves are most frequently affected.142 CNS manifestations occur more rarely and include cranial nerve impairment (especially optic neuritis), seizure, subarachnoid hemorrhage, and cerebral infarction. Skin, GI, renal, and other systemic alterations have been well described in EGPA. Skin findings are present in 40% to 70% of cases and may develop in localized crops. They can manifest as nonthrombocytopenic purpura (particularly on the lower extremities), urticaria, a maculopapular rash, petechiae, ecchymoses, or livedo reticularis. Skin biopsies typically demonstrate eosinophilic infiltration and leukocytoclastic vasculitis.161 Tender cutaneous or subcutaneous nodules (which may ulcerate) containing extravascular granulomas may preferentially involve the fingers, scalp, and extensor surface of the elbow.142 GI manifestations of EGPA are present in up to 60% of cases. They can include eosinophilic gastroenteritis or vasculitis that can lead to diarrhea, abdominal pain, intestinal obstruction, cholecystitis, pancreatitis, bleeding, liver function test abnormalities, and bowel perforation.142 GI disease also carries a poor prognosis142 and is the second leading cause of death in patients with EGPA. A degree of renal insufficiency occurs in 25% to 50% of patients with EGPA. Eosinophilic interstitial nephritis with necrotizing features is the most common histopathologic finding, but focal segmental glomerulosclerosis, hematuria, and proteinuria also occur.143,162,163 Severe, difficult-to-control hypertension is also a major sequela of EGPA (in 25%–75% of cases) and may be due to recurrent renal infarction. In contrast to granulomatosis with polyangiitis, overt renal failure is not commonly seen in EGPA.143 Mild lymphadenopathy (in 30%–40%), rheumatologic manifestations (migratory polyarthralgias, myalgias, temporal arteritis), urologic disease (ureteral, urethral, prostatic), and ocular manifestations (scleritis, uveitis, optic neuropathy, conjunctival nodules)142 have also been described.
The diagnosis of EGPA is based on clinical features with corroborating laboratory and/or histologic findings. There is no single laboratory test specific for a diagnosis of EGPA. A majority of patients with EGPA have a striking but fluctuating degree of peripheral blood eosinophilia (mean values between 5 and 20,000/mm3; 20% to 90% of the WBC differential) at diagnosis,1,142 generally greater than that seen with asthma alone. The degree of eosinophilia tends to parallel the vasculitis activity1 and may be suppressed by corticosteroid treatment of asthma. Serum C-reactive protein levels and total IgE levels are elevated (range, 500–1000 U/mL) in 75% of patients. Elevated exhaled breath condensate and BAL fluid concentrations of the eicosanoid 12-HETE can help distinguish EGPA from asthma without EGPA.164 Elevated blood levels of the Th2 T lymphocyte cytokines IL-4, IL-5, and IL-13165 as well as serum IgG4,166 TARC/CCL17,167 urinary eosinophil–derived neurotoxin,168 and serum eotaxin-3 levels169,170 also correlate with disease activity. BAL fluid of patients with active EGPA also contains increased levels of the Th2 cytokines IL-4, IL-5, and IL-10 as compared with inactive EGPA.171,172 Most patients have a normochromic, normocytic anemia, and moderate elevation of their ESR. Rheumatoid factor titers may be positive, but antinuclear antibodies are usually negative.142 Hypergammaglobulinemia and circulating immune complexes may also be seen.
Approximately 40% to 60% of patients have positive antinuclear cytoplasmic antibody with a perinuclear staining pattern (pANCA).159,160,173 The majority of these are directed against myeloperoxidase (MPO-ANCA) and a minority against proteinase 3 (PR3-ANCA). The absence of ANCA does not exclude the diagnosis. Indeed, recent case series suggest that persons positive for ANCA have different clinical features than those without.174 Persons with positive ANCA (the “vasculitic phenotype”) tend to manifest evidence of biopsy-proven vasculitis, purpura, ENT manifestations, peripheral neuropathy, and renal disease but have less cardiac involvement.159,160,175 In contrast, those without ANCA (the “eosinophilic tissue infiltration phenotype”) manifest predominantly with fever, eosinophilic pneumonia, and eosinophilic myocarditis.159,160,176
Laboratory examination of pleural fluid, if present, reveals an eosinophil-predominant exudate with low glucose levels.142 Pleural biopsy shows chronic pleuritis with eosinophilic infiltration. BAL reveals an increased percentage of eosinophils,177 the magnitude of which is generally less than that seen with CEP or idiopathic HES. However, patients have been described whose BAL fluid leukocyte differential contained 81% eosinophils. Electrocardiogram, echocardiography, N-terminal pro-brain natriuretic peptide and tropinin-1 levels, and/or cardiac MRI are recommended to evaluate patients with suspected EGPA for the presence of cardiac involvement.142,178,179 18FDG/13N ammonia positron emission tomography (PET) imaging may also be useful to identify cardiac involvement in EGPA. Magnetic resonance imaging may show T2-weighted signals in subcortical matter suggestive of CNS vasculitis.142
Although lung biopsy is rarely required for diagnosis, the histopathologic hallmarks of EGPA vary depending on the stage of illness but include tissue (interstitial, blood vessel, and alveolar) infiltration by eosinophils, eosinophilic necrotizing giant cell vasculitis of small arteries, arterioles, and, to a lesser extent, small veins, venules, and capillaries and perivascular and interstitial eosinophilic granulomas (typically microscopic).180–182 Both pulmonary and systemic vessels may be affected. The precise histopathology of vascular impairment depends on the stage of the lesion. Early lesions demonstrate eosinophilic infiltration of the vessels and perivascular region (Fig. 71-5). Later lesions are characterized by necrotizing arteritis or vessel obliteration and scarring. The extent of vascular impairment varies from mild, eosinophilic perivascular cuffing to severe transmural inflammation with necrotization. Lesions may be sparse or widespread. In the lung the allergic granulomas may have central necrosis with eosinophilic infammation, may involve interlobular septae and extend along the pleura. Diffuse capillaritis and alveolar hemmorhage has been reported. Eosinophilic lymphadenopathy may also be present. Biopsies of skin, nerve, or muscle may also confirm the diagnosis.156
Pathologic appearance of small arteriole in Churg–Strauss vasculitis. Intense perivascular inflammation with eosinophilia is present.
The pathogenesis of EGPA remains poorly understood. A possible pathogenic role of antineutrophil cytoplasmic antibody (ANCA) is suggested by the finding of ANCA in 40% to 60% of patients with EGPA.144 ANCA may contribute to tissue inflammation and injury by activation of inflammatory cells, release of proteolytic enzymes, and generation of oxidative stress142 but its presence may be a consequence rather than a cause of the tissue injury. As noted previously, the clinical features of EGPA differ between persons with versus without ANCA.159,160 The strong association with allergy, atopy, eosinophilia, and elevated blood levels of Th2 T lymphocyte-derived cytokines and IgE (especially during the vasculitic phase of the disease) has raised the likelihood of augmented Th2 immunity.142,172,183–185
Reductions in T regulatory cells183,186 and augmented Th1 immunity187 have also been demonstrated. Eosinophils likely also contribute significantly to the tissue injury188 since blood, BAL, and urine specimens contain eosinophil-derived cytotoxic granule proteins.142 Based on these collective findings, it has been proposed that repeated antigenic stimulation in patients with a heightened T-cell and eosinophil response may be important in the development of the disorder. Heightened humoral immunity with immune complex disease may also play a role. Genetic factors including polymorphisms in the IL-10 gene189 and HLA-DRB107 and HLA-DRB4146 may be important in predilection to develop EGPA.
The relationship between the pathophysiology of asthma in EGPA to that of asthma without EGPA also remains uncertain. CEP with asthma may precede EGPA in up to 50% of cases. A decrease in the number of blood regulatory T cells (Treg) occurs in persons who develop EGPA following CEP, but not in those with CEP who do not later develop EGPA or in persons with asthma without EGPA.186,190 This suggests that maintenance of normal numbers of Treg cells may protect against the development of EGPA. Also, a strong association has been noted between the use of leukotriene receptor antagonists (LTRA) and 5-lipoxygenase inhibitors,172,191–194 as well as other asthma therapies including inhaled glucocorticoids195,196 and omalizumab197 and the development of EGPA. These findings raise question as to whether these agents may serve as triggers for the disease. The appearance of EGPA following reduction in systemic corticosteroid dosing in many of these reports raises the likelihood that pre-existing, underlying EGPA that was being treated with corticosteroids is unmasked by the administration of these agents and the reduction in corticosteroid dose. Thus, it remains uncertain whether any of these agents may be causally related to the onset of EGPA. Patients with steroid-dependent asthma, in whom the diagnosis of EGPA has not been demonstrated or entertained, should be monitored closely for evidence of EGPA when steroid doses are tapered, or when symptoms escalate despite systemic corticosteroid treatment and require intensification of medical therapy. EGPA has also been reported following inhalational use of cocaine,198 following exposure to other medications, birds, vaccinations, and various infectious pathogens.13,142
The diagnosis of EGPA is generally based on clinical features. Various diagnostic criteria have been reported.148,182 In 1990, the American College of Rheumatology published diagnostic criteria for EGPA,182 based on assessments of the sensitivity and specificity of the diagnostic criteria used previously. The presence of at least four out of six of the following criteria yielded 85% sensitivity and 99.7% specificity in establishing the diagnosis: (1) asthma, (2) peripheral eosinophilia greater than 10%, (3) mono- or polyarthropathy, (4) migratory or transient pulmonary infiltrates, (5) paranasal sinus abnormality, and (6) extravascular eosinophils in a blood vessel on a biopsy specimen. The presence of asthma or allergy as well as more than 10% eosinophilia was 95% sensitive and 99% specific in distinguishing EGPA among a subgroup of patients with well-documented systemic vasculitis. Subsequently, the Chapel Hill Consensus Conference recommended that diagnostic criteria for EGPA include1 appropriate clinical setting and histopathology and2 eosinophil-rich and granulomatous inflammation involving the respiratory tract and necrotizing vasculitis affecting small and medium vessels with associated asthma and eosinophilia.199 However, these criteria require tissue biopsy and are less sensitive for EGPA than others that have been proposed; hence, they may be less useful to assist diagnosis in the routine clinical setting. Open lung biopsy is the gold standard for tissue biopsy but is rarely necessary. Transbronchial biopsy may reveal the diagnosis if there is alveolar involvement, but is often nondiagnostic. Biopsy of other sites (e.g., skin, pericardium, muscle, nerve, gut), with or without immunostaining, may assist in establishing the diagnosis in selected cases, although demonstration of characteristic histopathologic changes are not essential for establishing the diagnosis. The diagnosis may be elusive since EGPA may be suppressed by corticosteroid treatment for asthma.
The differential diagnosis of EGPA includes polyarteritis nodosa, microscopic angiitis, granulomatosis with polyangiitis, CEP, ABPA, idiopathic HES, Loeffler syndrome, asthma, fungal or parasitic infection, drug-induced vasculitis, sarcoidosis, and Hodgkin lymphoma. EGPA can be distinguished from granulomatosis with polyangiitis since compared with the latter, patients with EGPA have nasal polyps and allergic rhinitis but lack significant necrotizing upper airway lesions and cavitation of lung nodules, and are more likely to have pANCA (in contrast to the c-ANCA seen in granulomatosis with polyangiitis). Also, patients with EGPA are less likely to develop renal failure, and vasculitic neuropathy and asthma/eosinophilia are not typical features of granulomatosis with polyangiitis. EGPA can be distinguished from MPO-ANCA–positive microscopic angiitis since patients with the latter syndrome have leukocytoclastic vasculitis without granulomas and do not have upper airway involvement, asthma, and eosinophilia. Further, unlike EGPA, cardiac involvement is rare in MPO-ANCA–positive vasculitis. EGPA may be difficult to distinguish from idiopathic HES among persons who lack ANCA and/or neuropathy without overt vasculitis.200 EGPA is more likely among those with lower eosinophil counts who later develop vasculitis.
Patients in whom EGPA goes untreated have a poor prognosis; up to 50% die within 3 months after the onset of vasculitis. As such, efforts at early recognition and treatment are important.
Two randomized controlled trials of therapy for EGPA have been conducted recently.201,202 The choice of treatment depends in part on the clinical features at the time of diagnosis, since the presence of some features portends a worse prognosis and requires a more aggressive treatment approach.201–203 Poor prognostic features include age >65 years, cardiac, GI, CNS, and renal involvement (with serum creatinine >150 μmol/L) and absence of ENT involvement.175,203 In persons without poor prognostic features, corticosteroids are the mainstay of treatment and generally lead to dramatic clinical improvement, with disease stabilization or cure. Prednisone, 1 to 1.5 mg/kg/d (or 60 mg per day in adults) is given for 3 to 12 weeks, aiming to eliminate constitutional symptoms and cardiac, renal, neurologic, or other vasculitic manifestations. Higher doses (e.g., 15 mg/kg/d methylprednisolone) are occasionally required for control of life-threatening symptoms. Severe hypertension and mononeuritis multiplex often require prolonged steroid treatment and may be difficult to eliminate. Once the vasculitic phase is controlled, steroids may be tapered, with doses titrated to maintain disease control. Low-dose prednisone (e.g., 5–10 mg) is often given every day or every other day for up to 1 year. Although relapses are uncommon, patients should be followed closely for evidence of clinical deterioration, and should have periodic screening of total WBC and differential, ESR, and IgE levels. Most reports suggest the pANCA is not useful to monitor disease activity or direct therapeutic intervention, but one recent case series demonstrated that ANCA-positive persons had more frequent relapses and lower 5-year relapse-free survival rates than ANCA-negative persons.175 Treatment with cytotoxic immunosuppressive agents, such as cyclophosphamide or azathioprine, should be administered in patients whose condition fails to improve or have relapses despite steroid treatment.
Persons who have poor prognostic features at the time of disease presentation,142,204 or who have severe systemic involvement should receive high-dose intravenous methylprednisolone or oral prednisone for 3 days plus induction therapy with cyclophosphamide (2 mg/kg/d orally or 0.6–0.7 g/m2 intravenously day 1, 15, and 30 and then every 3 to 4 weeks theareafter).205 Patients treated with cyclophosphamide should be monitored closely for hemorrhagic cystitis, renal insufficiency, bone marrow suppression, bladder fibrosis, and urologic malignancies. Patients with severe disease treated with corticosteroids and cyclophosphamide have better survival than those treated with corticosteroids alone. Azathioprine maintenance therapy (2 mg/kg/d) for 18 to 24 months may be beneficial once remission of disease has been achieved.142,143,202
Intravenous immunoglobulin (IvIg) may be beneficial for use in pregnant women and for reducing symptoms and organ involvement and improving long-term disease control among persons with severe organ involvement. The number of blood Treg cells was increased among persons with EGPA treated with Iv Ig and conventional therapy.206 The anti-IL5 antibody mepolizumab has also demonstrated efficacy as a steroid-sparing agent in small uncontrolled case series.207,208 The immunoregulatory cytokine interferon-α (IFN-α) has led to improved pulmonary function tests, reduction in corticosteroid dose, and decreased WBC count and may be considered as another alternative treatment in persons with refractory disease209 but may be of limited efficacy and can cause cardiac toxicity. Plasma exchange may also be a successful adjunct treatment in some patients,142,210 particularly those with ANCA positivity and glomerulonephritis. Finally, rituximab may be effective as an alternate therapeutic agent but further studies are needed to clarify benefits of treatment in EGPA as compared with other forms of ANCA-positive vasculitis.211 β-Blockers should be avoided in the management of EGPA-related hypertension, owing to the risk of bronchospasm and CHF. Persons undergoing immunosuppressive treatment for EGPA should receive prophylactic treatment to prevent infection with pneumocystis jirovecii.
Prolonged treatment may be necessary to maintain disease control. Long-term overall remission can be achieved in approximately 81% to 92% of patients142,156; relapses occur in 25% of cases and are most common within 1 year.142 In a series of 30 patients collected over the period 1950 to 1974, a median survival of more than 9 years was reported in patients treated with steroids; 1-year survival was 90%, 3-year survival was 76%, and 62% survival was noted at 5 years. More recent studies suggest 85% to 100% survival at 5 years.1,175,205